<--back--   home   --next-->

There are many directions of current and future work. Since the amount of deflection is related to the anisotropy, it might be possible to use a hue-ball type method to assign opacity as well as color. Conversely, it would be interesting to use the barycentric space to assign color as well as opacity. There are some unanswered questions about the best way to interpolate tensors. We did it by interpolating component-wise, but perhaps there's something more intelligent. Also, the colors on the hue-ball did not incorporate any knowledge about perceptually uniform color spaces, and perhaps they should. Finally, it would be interesting to apply any or all of these techniques to non-medical data.